Fix ARM trailing data directive size

[floze-the-genius]

Summary

• initialize ARM mapping symbols before inferring zero-sized function spans
• preserve terminal word-sized $d data directives while retaining trailing halfword padding removal
• add the issue ARM/Thumb: Final data directive size is incorrect #382 object fixture and regression test

Testing

• cargo test

Closes#382

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:d69d306dfb

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:e2b98eb517

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:2c6124bd9f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[floze-the-genius]

@LagoLunatic Thanks for supplying the corrected object fixture on #382. The three edge cases raised during automated review are now covered and all threads are resolved: 4n+2 padding, overlapping relocations in the final word, and data mappings starting at 2 mod 4. Current head 64cf492 is mergeable; 19 checks pass (4 release-only jobs are skipped). Could you review whether the ARM size-inference behavior now matches the intended mapping-symbol semantics? I’m happy to rebase onto current main if you prefer that before review.